U.S. patent application number 14/748148 was filed with the patent office on 2015-12-24 for system and method for partition migration in a multitenant application server environment.
The applicant listed for this patent is ORACLE INTERNATIONAL CORPORATION. Invention is credited to TRILOK AGARWAL, JOSEPH DIPOL, NAZRUL ISLAM, MARTIN MARES, RAJIV MORDANI, CHANDA PATEL.
Application Number | 20150372938 14/748148 |
Document ID | / |
Family ID | 53511007 |
Filed Date | 2015-12-24 |
United States Patent
Application |
20150372938 |
Kind Code |
A1 |
PATEL; CHANDA ; et
al. |
December 24, 2015 |
SYSTEM AND METHOD FOR PARTITION MIGRATION IN A MULTITENANT
APPLICATION SERVER ENVIRONMENT
Abstract
In accordance with an embodiment, described herein is a system
and method for partition migration in a multitenant application
server environment. Each resource group, or partition which
includes that resource group, can be associated with a virtual
target (e.g., virtual host) information that defines a target to
which that resource group or partition should be targeted and
deployed. A web tier component (e.g., Oracle Traffic Director)
includes a routing information, which maps the virtual target for a
partition to a corresponding target such as a server or cluster, so
that requests for that partition are directed to the appropriate
target, for example as part of one or more sessions. When a
partition is migrated, session information is replicated between an
original target and a new target, and the web tier component
provided with a revised routing information, so that subsequent
requests for that partition are directed to the new target.
Inventors: |
PATEL; CHANDA; (Milpitas,
CA) ; MORDANI; RAJIV; (Sunnyvale, CA) ; ISLAM;
NAZRUL; (Santa Clara, CA) ; DIPOL; JOSEPH;
(Sunnyvale, CA) ; MARES; MARTIN; (Prague, CZ)
; AGARWAL; TRILOK; (Santa Clara, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
ORACLE INTERNATIONAL CORPORATION |
Redwood Shores |
CA |
US |
|
|
Family ID: |
53511007 |
Appl. No.: |
14/748148 |
Filed: |
June 23, 2015 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62016035 |
Jun 23, 2014 |
|
|
|
62055106 |
Sep 25, 2014 |
|
|
|
Current U.S.
Class: |
709/226 |
Current CPC
Class: |
H04L 67/10 20130101;
G06F 9/5077 20130101; H04L 47/70 20130101; H04L 45/04 20130101;
G06F 9/5088 20130101 |
International
Class: |
H04L 12/911 20060101
H04L012/911; H04L 12/715 20060101 H04L012/715; H04L 29/08 20060101
H04L029/08 |
Claims
1. A system for supporting partition migration in an application
server environment, comprising: one or more computers, including an
application server environment having a domain for execution of
software applications, and one or more partitions and resource
groups, wherein each partition provides an administrative and
runtime subdivision of the domain and can include one or more
resource groups, and wherein each resource group is a collection of
deployable applications or resources that can reference a resource
group template; and wherein each resource group, or the partition
which includes that resource group, can be associated with a
virtual target information that defines a target to which that
resource group or partition should be targeted and deployed, and
wherein a web tier component includes a routing information, which
maps the virtual target for a particular partition to a
corresponding target such as a server or cluster, so that requests
for that partition are directed to the appropriate target.
2. The system of claim 1, wherein the application server or a cloud
environment includes a Java EE application server.
3. The system of claim 1, wherein when a particular partition is
migrated, session information is replicated between an original
target and a new target, and the web tier component provided with a
revised routing information, so that subsequent requests for that
partition are directed to the new target.
4. The system of claim 1, wherein the system is provided within a
cloud environment, to support multiple tenants operating within the
cloud environment, including that the system can optionally
associate one or more partitions with a tenant, for use by the
tenant.
5. A method for supporting partition migration in an application
server environment, comprising: providing, at one or more
computers, an application server environment having a domain for
execution of software applications, and one or more partitions and
resource groups, wherein each partition provides an administrative
and runtime subdivision of the domain and can include a resource
group, and wherein each resource group is a collection of
deployable applications or resources that can reference a resource
group template; and wherein each resource group, or the partition
which includes that resource group, can be associated with a
virtual target information that defines a target to which that
resource group or partition should be targeted and deployed, and
wherein a web tier component includes a routing information, which
maps the virtual target for a particular partition to a
corresponding target such as a server or cluster, so that requests
for that partition are directed to the appropriate target.
6. The method of claim 5, wherein the application server or a cloud
environment includes a Java EE application server.
7. The method of claim 5, wherein when a particular partition is
migrated, session information is replicated between an original
target and a new target, and the web tier component provided with a
revised routing information, so that subsequent requests for that
partition are directed to the new target.
8. The method of claim 5, wherein the method is performed within a
cloud environment, to support multiple tenants operating within the
cloud environment, including that the system can optionally
associate one or more partitions with a tenant, for use by the
tenant.
9. A non-transitory computer readable storage medium, including
instructions stored thereon which when read and executed by one or
more computers cause the one or more computers to perform the steps
comprising: providing an application server environment having a
domain for execution of software applications, and one or more
partitions and resource groups, wherein each partition provides an
administrative and runtime subdivision of the domain and can
include a resource group, and wherein each resource group is a
collection of deployable applications or resources that can
reference a resource group template; and wherein each resource
group, or the partition which includes that resource group, can be
associated with a virtual target information that defines a target
to which that resource group or partition should be targeted and
deployed, and wherein a web tier component includes a routing
information, which maps the virtual target for a particular
partition to a corresponding target such as a server or cluster, so
that requests for that partition are directed to the appropriate
target.
10. The non-transitory computer readable storage medium of claim 9,
wherein the application server or a cloud environment includes a
Java EE application server.
11. The non-transitory computer readable storage medium of claim 9,
wherein when a particular partition is migrated, session
information is replicated between an original target and a new
target, and the web tier component provided with a revised routing
information, so that subsequent requests for that partition are
directed to the new target.
12. The non-transitory computer readable storage medium of claim 9,
wherein the application server environment is provided within a
cloud environment, to support multiple tenants operating within the
cloud environment, including that the system can optionally
associate one or more partitions with a tenant, for use by the
tenant.
Description
CLAIM OF PRIORITY AND CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of priority to U.S.
Provisional Application titled "SYSTEM AND METHOD FOR PARTITION
MIGRATION IN A MULTITENANT APPLICATION SERVER ENVIRONMENT",
Application No. 62/016,035, filed Jun. 23, 2014; and U.S.
Provisional Application titled "SYSTEM AND METHOD FOR PARTITION
MIGRATION IN A MULTITENANT APPLICATION SERVER ENVIRONMENT",
Application No. 62/055,106, filed Sep. 25, 2014; and is related to
U.S. Patent Application titled "SYSTEM AND METHOD FOR SUPPORTING
MULTI-TENANCY IN AN APPLICATION SERVER, CLOUD, OR OTHER
ENVIRONMENT", Application Ser. No. 14/601,883, filed Jan. 21, 2015;
each of which above applications are herein incorporated by
reference.
COPYRIGHT NOTICE
[0002] A portion of the disclosure of this patent document contains
material which is subject to copyright protection. The copyright
owner has no objection to the facsimile reproduction by anyone of
the patent document or the patent disclosure, as it appears in the
Patent and Trademark Office patent file or records, but otherwise
reserves all copyright rights whatsoever.
FIELD OF INVENTION
[0003] Embodiments of the invention are generally related to
application servers and cloud environments, and are particularly
related to a system and method for partition migration in a
multitenant application server environment.
BACKGROUND
[0004] Application servers generally provide a managed environment
within which software applications can be deployed and run.
Cloud-based environments allow applications to run within and take
advantage of the distributed resources provided by a cloud. Such
environments can support large numbers of users or tenants, some of
which may have specific requirements that are particular to that
user or tenant. These are some examples of the type of environments
in which embodiments of the present invention can be used.
SUMMARY
[0005] In accordance with an embodiment, described herein is a
system and method for partition migration in a multitenant
application server environment. Each resource group, or the
partition which includes that resource group, can be associated
with a virtual target (e.g., virtual host) information that defines
a target to which that resource group or partition should be
targeted and deployed. A web tier component (e.g., Oracle Traffic
Director) includes a routing information, which maps the virtual
target for a particular partition to a corresponding target such as
a server or cluster, so that requests for that partition are
directed to the appropriate target, for example as part of one or
more sessions. When a particular partition is migrated, session
information is replicated between an original target and a new
target, and the web tier component provided with a revised routing
information, so that subsequent requests for that partition are
directed to the new target.
BRIEF DESCRIPTION OF THE FIGURES
[0006] FIG. 1 illustrates a system for supporting multi-tenancy in
an application server, cloud, or other environment, in accordance
with an embodiment.
[0007] FIG. 2 further illustrates a system for supporting
multi-tenancy in an application server, cloud, or other
environment, in accordance with an embodiment.
[0008] FIG. 3 further illustrates a system for supporting
multi-tenancy in an application server, cloud, or other
environment, in accordance with an embodiment.
[0009] FIG. 4 illustrates a domain configuration for use with an
exemplary multi-tenant environment, in accordance with an
embodiment.
[0010] FIG. 5 further illustrates an exemplary multi-tenant
environment, in accordance with an embodiment.
[0011] FIG. 6 illustrates the use of partition migration in an
application server, cloud, or other environment, in accordance with
an embodiment.
[0012] FIG. 7 further illustrates the use of partition migration in
an application server, cloud, or other environment, in accordance
with an embodiment.
[0013] FIG. 8 illustrates a method of using partition migration in
an application server, cloud, or other environment, in accordance
with an embodiment.
DETAILED DESCRIPTION
[0014] In accordance with an embodiment, described herein is a
system and method for partition migration in a multitenant
application server environment. Each resource group, or the
partition which includes that resource group, can be associated
with a virtual target (e.g., virtual host) information that defines
a target to which that resource group or partition should be
targeted and deployed. A web tier component (e.g., Oracle Traffic
Director) includes a routing information, which maps the virtual
target for a particular partition to a corresponding target such as
a server or cluster, so that requests for that partition are
directed to the appropriate target, for example as part of one or
more sessions. When a particular partition is migrated, session
information is replicated between an original target and a new
target, and the web tier component provided with a revised routing
information, so that subsequent requests for that partition are
directed to the new target.
Application Server (e.g., Multi-Tenant, MT) Environment
[0015] FIG. 1 illustrates a system for supporting multi-tenancy in
an application server, cloud, or other environment, in accordance
with an embodiment.
[0016] As illustrated in FIG. 1, in accordance with an embodiment,
an application server (e.g., multi-tenant, MT) environment 100, or
other computing environment which enables the deployment and
execution of software applications, can be configured to include
and operate according to a domain 102 configuration that is used at
runtime to define an application server domain.
[0017] In accordance with an embodiment, the application server can
include one or more partitions 104 that are defined for use at
runtime. Each partition can be associated with a globally unique
partition identifier (ID) and partition configuration, and can
further include one or more resource groups 124, together with a
reference to a resource group template 126 and/or
partition-specific applications or resources 128. Domain-level
resource groups, applications and/or resources 140 can also be
defined at the domain level, optionally with a reference to a
resource group template.
[0018] Each resource group template 160 can define one or more
applications A 162, B 164, resources A 166, B 168, and/or other
deployable applications or resources 170, and can be referenced by
a resource group. For example, as illustrated in FIG. 1, resource
group 124 in partition 104 can reference 190 resource group
template 160.
[0019] Generally, a system administrator can define partitions,
domain-level resource groups and resource group templates, and
security realms; while a partition administrator can define aspects
of their own partition, for example, by creating partition-level
resource groups, deploying applications to the partition, or
referencing specific realms for the partition.
[0020] FIG. 2 further illustrates a system for supporting
multi-tenancy in an application server, cloud, or other
environment, in accordance with an embodiment.
[0021] As illustrated in FIG. 2, in accordance with an embodiment,
a partition 202 can include, for example, a resource group 205
which includes a reference 206 to a resource group template 210, a
virtual target (e.g., virtual host) information 207, and a
pluggable database (PDB) information 208. A resource group template
(e.g., 210) can define, for example, a plurality of applications A
211 and B 212, together with resources such as a Java Message
Server (JMS) server 213, store-and-forward (SAF) agent 215, mail
session component 216, or Java Database Connectivity (JDBC)
resource 217.
[0022] The resource group template illustrated in FIG. 2 is
provided by way of example; in accordance with other embodiments,
different types of resource group templates and elements can be
provided.
[0023] In accordance with an embodiment, when a resource group
within a partition (e.g., 202) references 220 a particular resource
group template (e.g., 210), information associated with a
particular partition can be used in combination with the referenced
resource group template, to indicate a partition-specific
information 230, for example a partition-specific PDB information.
The partition-specific information can then be used by the
application server to configure resources, for example a PDB
resource, for use by the partition. For example, partition-specific
PDB information associated with partition 202 can be used, by the
application server, to configure 232 a container database (CDB) 236
with an appropriate PDB 238, for use by that partition.
[0024] Similarly, in accordance with an embodiment, a virtual
target information associated with a particular partition can be
used to define 239 a partition-specific virtual target 240, for use
by the partition, e.g., baylandurgentcare.com, which can then be
made accessible via a uniform resource locator (URL), e.g.,
http://baylandurgentcare.com.
[0025] FIG. 3 further illustrates a system for supporting
multi-tenancy in an application server, cloud, or other
environment, in accordance with an embodiment.
[0026] In accordance with an embodiment, a system configuration
such as a config.xml configuration file, is used to define a
partition, including configuration elements for resource groups
associated with that partition, and/or other partition properties.
Values can be specified per-partition using property name/value
pairs.
[0027] In accordance with an embodiment, a plurality of partitions
can be executed within a managed server/cluster 242, or a similar
environment which can provide access to a CDB 243, and which are
accessible via a web tier 244. This allows, for example, a domain
or partition to be associated with one or more of the PDBs (of the
CDB).
[0028] In accordance with an embodiment, each of the plurality of
partitions, in this example partition A 250 and partition B 260,
can be configured to include a plurality of resources associated
with that partition. For example, partition A can be configured to
include a resource group 251 which contains an application A1 252,
application A2 254, and JMS A 256, together with a datasource A 257
associated with PDB A 259, wherein the partition is accessible via
a virtual target A 258. Similarly, partition B 260 can be
configured to include a resource group 261 which contains an
application B1 262, application B2 264, and JMS B 266, together
with a datasource B 267 associated with PDB B 269, wherein the
partition is accessible via a virtual target B 268.
[0029] While several of the above examples illustrate use of CDB
and PDBs, in accordance with other embodiments, other types of
multi-tenant or non-multi-tenant databases can be supported,
wherein a particular configuration can be provided for each
partition, for example through the use of schemas, or the use of
different databases.
Resources
[0030] In accordance with an embodiment, a resource is a system
resource, application, or other resource or object that can be
deployed to a domain of the environment. For example, in accordance
with an embodiment, a resource can be an application, JMS, JDBC,
JavaMail, WLDF, data source, or other system resource or other type
of object that can be deployed to a server, cluster, or other
application server target.
Partitions
[0031] In accordance with an embodiment, a partition is a runtime
and administrative subdivision or slice of a domain that can be
associated with a partition identifier (ID) and configuration, and
can contain applications and/or refer to domain-wide resources
through the use of resource groups and resource group
templates.
[0032] Generally, a partition can contain its own applications,
refer to domain wide applications via resource group templates, and
have its own configuration. Partitionable entities can include
resources, for example JMS, JDBC, JavaMail, WLDF resources, and
other components, such as JNDI namespace, network traffic, work
managers, and security policies and realms. In the context of a
multi-tenant environment, the system can be configured to provide
tenant access to the administrative and runtime aspects of
partitions associated with a tenant.
[0033] In accordance with an embodiment, each resource group within
a partition can optionally reference a resource group template. A
partition can have multiple resource groups, and each of them can
reference a resource group template. Each partition can define
properties for configuration data that is not specified in the
resource group templates to which the partition's resource groups
refer. This enables the partition to act as a binding of deployable
resources defined in a resource group template, to specific values
for use with that partition. In some cases, a partition can
override configuration information specified by a resource group
template.
[0034] In accordance with an embodiment, a partition configuration,
as defined for example by a config.xml configuration file, can
include a plurality of configuration elements, for example:
"partition", which contains the attributes and child elements that
define the partition; "resource-group", which contains the
applications and resources deployed to the partition;
"resource-group-template", which contains applications and
resources defined by that template;
"jdbc-system-resource-override", which contains a database-specific
service name, user name, and password; and "partition-properties",
which contains property key values that can be used for macro
replacement in resource group templates.
[0035] Upon startup, the system can use the information provided by
the configuration file to generate partition-specific configuration
elements for each resource, from the resource group template.
Resource Groups
[0036] In accordance with an embodiment, a resource group is a
named, fully-qualified collection of deployable resources that can
be defined either at a domain or partition level, and can reference
a resource group template. The resources in a resource group are
considered fully-qualified in that the administrator has provided
all of the information needed to start or connect to those
resources, for example the credentials for connecting to a data
source, or the targeting information for a application.
[0037] A system administrator can declare resource groups at the
domain level, or at the partition level. At the domain level, a
resource group provides a convenient way to group related
resources. The system can manage resources declared in a
domain-level resource group the same as ungrouped resources, so
that the resources can be started during system start-up, and
stopped during system shut-down. An administrator can also stop,
start, or remove a resource in a group individually, and can act on
all the resources in a group implicitly by operating on the group.
For example, stopping a resource group stops all of the resources
in the group that are not already stopped; starting the resource
group starts any resources in the group that are not already
started; and removing the resource group removes all of the
resources contained in the group.
[0038] At the partition level, a system or partition administrator
can configure zero or more resource groups in a partition, subject
to any security restrictions. For example, in a SaaS use case,
various partition-level resource groups can refer to domain-level
resource group templates; while in a PaaS use case, partition-level
resource groups can be created that do not refer to resource group
templates, but instead represent applications and their related
resources that are to be made available within that partition
only.
[0039] In accordance with an embodiment, resource grouping can be
used to group together applications and the resources they use as a
distinct administrative unit within the domain. For example, in the
medical records (MedRec) application described below, a resource
grouping defines the MedRec application and its resources. Multiple
partitions can run the same MedRec resource group, each using a
partition-specific configuration information, such that the
applications that are part of each MedRec instance are made
specific to each partition.
Resource Group Templates
[0040] In accordance with an embodiment, a resource group template
is a collection of deployable resources that are defined at a
domain level, that can be referenced from a resource group, and
some of the information required to activate its resources may not
be stored as part of the template itself, such that it supports the
specification of partition level configuration. A domain can
contain any number of resource group templates, each of which can
include, for example, one or more related Java applications and the
resources on which those applications depend. Some of the
information about such resources may be the same across all
partitions, while other information may vary from one partition to
the next. Not all configuration need be specified at the domain
level--partition level configuration can instead be specified in
the resource group template through the use of macros, or property
name/value pairs.
[0041] In accordance with an embodiment, a particular resource
group template can be referenced by one or more resource groups.
Generally, within any given partition, a resource group template
can be referenced by one resource group at a time, i.e., not
simultaneously by multiple resource groups within the same
partition; however, it can be referenced at the same time by
another resource group in a different partition. The object
containing the resource group, e.g., the domain or partition, can
use property name/value assignments to set the value of any tokens
in the resource group template. When the system activates a
resource group template using a referencing resource group, it can
replace those tokens with values set in the resource group's
containing object. In some cases, the system can also use
statically-configured resource group templates and partitions to
generate runtime configuration for each partition/template
combination.
[0042] For example, in a SaaS use case, the system can activate the
same applications and resources multiple times, including once for
each partition that will use them. When an administrator defines a
resource group template they can use tokens to represent the
information that will be supplied elsewhere. For example, the
username to use in connecting to a CRM-related data resource can be
indicated in the resource group template as
\${CRMDataUsername}.
Tenants
[0043] In accordance with an embodiment, in a multi-tenant
environment, such as a multi-tenant (MT) application server
environment, a tenant is an entity that can be represented by, or
otherwise associated with, one or more partitions and/or one or
more tenant-aware applications.
[0044] For example, tenants can represent distinct user
organizations, such as different external companies, or different
departments within a particular enterprise (e.g., HR and Finance
departments), each of which can be associated with a different
partition. A tenant globally unique identity (tenant ID) is the
association of a particular user, at a particular moment in time,
with a particular tenant. The system can derive which tenant a
particular user belongs to from the user identity, for example by
referring to a user identity store. The user identity enables the
system to enforce those actions that a user is authorized to
perform, including, but not limited to, which tenant the user may
belong.
[0045] In accordance with an embodiment, the system enables
isolation of the administration and runtime of different tenants
from each other. For example, tenants can configure some behaviors
of their applications, and resources to which they have access. The
system can ensure that a particular tenant cannot administer
artifacts belonging to another tenant; and, at runtime, that the
applications working on behalf of a particular tenant refer only to
resources associated with that tenant, and not to resources
associated with other tenants.
[0046] In accordance with an embodiment, a tenant-unaware
application is one that contains no logic dealing with tenants
explicitly, such that any resources that the application uses may
be accessible regardless of what user submitted a request to which
the application is responding. In contrast, a tenant-aware
application includes logic that explicitly deals with tenants. For
example, based on a user's identity the application can derive the
tenant to which the user belongs and use that information to access
tenant-specific resources.
[0047] In accordance with an embodiment, the system enables users
to deploy applications that are explicitly written to be
tenant-aware, so that application developers can obtain the tenant
ID of a current tenant. The tenant-aware application can then use
the tenant ID to handle multiple tenants that are using a single
instance of the application.
[0048] For example, the MedRec application, which supports a single
doctor's office or hospital, can be exposed to two different
partitions or tenants, e.g., a Bayland Urgent Care tenant, and a
Valley Health tenant, each of which is able to access separate
tenant-specific resources, such as separate PDBs, without changing
the underlying application code.
Exemplary Domain Configuration and Multi-Tenant Environment
[0049] In accordance with an embodiment, applications can be
deployed to a resource group template at the domain level, or to a
resource group that is scoped to a partition or scoped to the
domain. Application configuration can be overridden using
deployment plans specified per-application, or per-partition.
Deployment plans can also be specified as part of a resource
group.
[0050] FIG. 4 illustrates a domain configuration for use with an
exemplary multi-tenant environment, in accordance with an
embodiment.
[0051] In accordance with an embodiment, when the system starts a
partition, it creates virtual targets (e.g., virtual hosts) and
connection pools, including one for each partition, to respective
database instances, according to the provided configuration.
[0052] Typically, each resource group template can include one or
more related applications and the resources on which those
applications depend. Each partition can provide configuration data
that is not specified in the resource group templates to which it
refers, by providing a binding of the deployable resources in
resource group templates to specific values associated with the
partition; including, in some cases, overriding certain
configuration information specified by the resource group template.
This enables the system to activate an application represented by a
resource group template differently for each partition, using the
property values each partition has defined.
[0053] In some instances, a partition may contain resource groups
that do not refer to resource group templates, or that directly
define their own partition-scoped deployable resources.
Applications and data sources that are defined within a partition
are generally available only to that partition. Resources can be
deployed so that they can be accessed from across partitions using
partition:<partitionName>/<resource JNDI name>, or
domain:<resource JNDI name>.
[0054] For example, a MedRec application can include a plurality of
Java applications, a data source, a JMS server, and a mail session.
To run the MedRec application for multiple tenants, the system
administrator can define a single MedRec resource group template
286, declaring those deployable resources in the template.
[0055] In contrast to domain-level deployable resources, the
deployable resources declared in a resource group template may not
be fully configured in the template, or cannot be activated as-is,
since they lack some configuration information.
[0056] For example, the MedRec resource group template may declare
a data source used by the applications, but it may not specify a
URL for connecting to the database. Partitions associated with
different tenants, for example, partition BUC-A 290 (Bayland Urgent
Care, BUC) and partition VH-A 292 (Valley Health, VH) can reference
one or more resource group templates, by each including a MedRec
resource group 293, 294 that references 296, 297 the MedRec
resource group template. The reference can then be used to create
302, 306, the virtual targets/virtual hosts for each tenant,
including a virtual host baylandurgentcare.com 304 associated with
the BUC-A partition, for use by the Bayland Urgent Care tenant; and
a virtual host valleyhealth.com 308 associated with the VH-A
partition, for use by the Valley Health tenant.
[0057] FIG. 5 further illustrates an exemplary multi-tenant
environment, in accordance with an embodiment. As illustrated in
FIG. 5, and continuing the example from above, in which two
partitions reference the MedRec resource group template, in
accordance with an embodiment, a servlet engine 310 can be used to
support a plurality of tenant environments, in this example a
Bayland Urgent Care Physician tenant environment 320, and a Valley
Health Physician tenant environment 330.
[0058] In accordance with an embodiment, each partition 321, 331
can define a different virtual target on which to accept incoming
traffic for that tenant environment, and a different URL 322, 332
for connecting to the partition and to its resources 324, 334,
including in this example either a bayland urgent care database, or
a valley health database respectively. The database instances can
use compatible schemas, since the same application code will
execute against both databases. When the system starts the
partitions, it can create the virtual targets and connection pools
to the respective database instances.
Partition Migration
[0059] In accordance with an embodiment, described herein is a
system and method for partition migration in a multitenant
application server environment. Each resource group, or the
partition which includes that resource group, can be associated
with a virtual target (e.g., virtual host) information that defines
a target to which that resource group or partition should be
targeted and deployed. A web tier component (e.g., Oracle Traffic
Director) includes a routing information, which maps the virtual
target for a particular partition to a corresponding target such as
a server or cluster, so that requests for that partition are
directed to the appropriate target, for example as part of one or
more sessions. When a particular partition is migrated, session
information is replicated between an original target and a new
target, and the web tier component provided with a revised routing
information, so that subsequent requests for that partition are
directed to the new target.
[0060] In accordance with an embodiment, partition migration can be
used to provide a seamless migrations, including preserving the
state of sessions associated with a partition when the partition is
migrated from one target to another target. This can be useful, for
example, in ensuring that a customer application is not disrupted
due to problems with the underlying computing infrastructure.
[0061] FIG. 6 illustrates the use of partition migration in an
application server, cloud, or other environment, in accordance with
an embodiment. As shown in FIG. 6, in accordance with an
embodiment, a first or original target cluster A 402 can include
one or more partitions including a particular partition A 404,
which is associated with a partition configuration 406 and one or
more partition-specific resource groups, applications or resources
408.
[0062] Optionally, the partition can reference a resource group
template 410.
[0063] Each resource group, or the partition which includes that
resource group, can be associated with a virtual target (e.g.,
virtual host) information that defines a target to which that
resource group or partition should be targeted and deployed, such
as virtual target A 420. A web tier component (e.g., a traffic
director, such as Oracle Traffic Director) 430 includes a routing
information 432, which maps the virtual target for a particular
partition to a corresponding target such as a server or cluster, so
that requests for that partition are directed to the appropriate
target, for example as part of one or more sessions 434.
[0064] As further illustrated, the system can include a second or
new target cluster B 442, to which partitions can be deployed.
[0065] FIG. 7 further illustrates the use of partition migration in
an application server, cloud, or other environment, in accordance
with an embodiment. As shown in FIG. 7, in accordance with an
embodiment, an indication can be received to migrate 444 a
particular partition, and any sessions currently operating thereon,
from the original target to the new target. Generally, the system
will only migrate partitions between targets that are associated
with a particular virtual target, for example between different
cluster members that are associated with a particular virtual
target (and not between different virtual targets).
[0066] In response, the particular partition is migrated, including
that session information is replicated between the original target
and the new target, and the web tier component is provided with a
revised routing information 446, which maps the virtual target for
the particular partition to the new target, so that subsequent
requests for that partition are directed 448 to the new target.
[0067] For example, in accordance with an embodiment, the
application server environment can inform the traffic director
(e.g., OTD), that a particular computer cluster will be going down
soon (say, for maintenance or failover reasons), and instruct the
traffic director to start sending requests to the expected new
cluster. In the meantime, the application server can replicate the
session information to the new cluster. Once the session
replication is complete, the application server environment can
inform the traffic director to remove the downgraded computer
cluster from its configuration, which can then also be removed from
any resource group targeting.
[0068] In accordance with an embodiment, when clusters are used as
targets, the system performs live migrations of sessions. When a
computer server is used as a target, the live migration
functionality may be disabled and a user-controlled manual
migration may be used, so the user can be informed of any possible
downtime or loss of session information.
[0069] FIG. 8 illustrates a method of using partition migration in
an application server, cloud, or other environment, in accordance
with an embodiment. As shown in FIG. 8, in accordance with an
embodiment, at step 460, an application server environment is
provided, including a domain for execution of software
applications, and one or more partitions and resource groups,
wherein each partition provides an administrative and runtime
subdivision of the domain and can include one or more resource
groups, and wherein each resource group is a collection of
deployable applications or resources that can reference a resource
group template.
[0070] At step 462, each resource group, or partition which
includes that resource group, can be associated with a virtual
target (e.g., virtual host) information that defines a target to
which that resource group or partition should be targeted and
deployed.
[0071] At step 464, an indication is received to migrate a
particular partition, and any sessions currently operating thereon,
from an original target to a new target.
[0072] At step 466, the particular partition is migrated, including
that session information is replicated between the original target
and the new target, and a web tier component (e.g., a traffic
director, such as Oracle Traffic director) is provided with a
revised routing information, which maps the virtual target for the
particular partition to the new target, so that subsequent requests
for that partition are directed to the new target.
Example Implementation
[0073] Additional details describing an exemplary embodiment are
provided below for purposes of illustration.
Live Migration Overview (For http and t3 Traffic)
[0074] In accordance with an embodiment, during a live migration,
the resource group is running. The Administrator initiates resource
group migration using, e.g., a JMX command and specifies the
virtual target mbean, existing target, and new target as
parameters. The new target is added to virtual target and the
resources are deployed to the new target. If a validation error is
thrown then the operation is terminated and the user is informed to
shut down the source resource group and perform a non-live
migration. If the current and new target is a cluster, the
clustering APIs for session replication and T3-re-routing are
initiated. The LCM APIs for OTD re-routing are initiated which
internally determines if OTD is front-ending this resource
group(s). Graceful shutdown of applications/resources targeted to
old target is initiated using resource group graceful shutdown for
specified target. The mbean configuration for virtual target is
updated to remove the old target.
Non-Live Migration Overview (For http and t3 Traffic)
[0075] In accordance with an embodiment, during a non-live
migration, the resource group is not running. The WLS Administrator
initiates resource group migration using, e.g., a JMX command and
specifies the virtual target mbean, existing target, new target as
parameters. The old target is removed from virtual target and new
target is added to the virtual target which internally deploys the
resources/apps to the new target. The LCM APIs for OTD update are
initiated which internally determines if OTD is front-ending this
resource group(s).
Resource Group Migration API
[0076] In accordance with an embodiment, migration will be
initiated using new service PortablePartitionManagerMbean. [0077]
ResourceGroupMigrationTask=migrateResourceGroup(TargetMBean
VirtualTarget, TargetMBean currentTarget, [0078] TargetMBean
newTarget, long timeout);
[0079] A validation exception will be thrown if a new migration
call is initiated while another migration call is in process for
the same virtual target. It is not possible to start two
simultaneous migration processes in one resource group.
ResourceGroupMigrationTask contains information about state of this
long running process.
Adding New Target to Virtual Target
[0080] In accordance with an embodiment, in the live migration case
the new target is added to the virtual target and the application
and resources in the RG are deployed on the new target. If any
application or resource does not support multi-targeting then they
will throw a validation exception. In this case the migration
operation will terminate reporting that the resource group does not
support live migration, and that the RG must be shutdown first to
do a non-live migration.
Session Replication and T3-Rerouting
[0081] In accordance with an embodiment, after the new target is
added, an WLS-MT clustering session replication feature can be used
to synchronize sessions between old cluster and new cluster, in
which session states are copied to some servers in new
topology--new Target. After that callback event is send but process
continue to synchronize changed sessions. The migration clustering
subsystem notifies all connected t3 clients about target change. T3
clients will reconnect to new urls from the target and continue its
communication transparently with migrated application. In the case
the entire cluster should go down the client should have the
ability to reach a well-known address and rediscover a new cluster
to route to.
HTTP Traffic Re-Routing Using WLS-MT-Lifecycle
[0082] In accordance with an embodiment, migration functionality
can leverage the availability of a lifecycle module for migration,
which attempts to provide configuration integration across
components which were developed independently, in different
languages, and using different architectures for configuration and
administration.
Seamless Case
[0083] In accordance with an embodiment, during seamless cases, the
application server will invoke the LCM with a partition migration
REST call: [0084] //migrate all partitions associated with the
environment
POST/management/lifecycle/latest/environments/{name}l/partitions/{id}/mig-
rate
[0085] The payload will contain the details of new cluster where
the partition is deployed and phase: start-migrate which would in
turn invoke the OTD plugin for configuration change in OTD.
Following are the steps happening on OTD side as specified in OTD
functional specification.
[0086] When a partition migration is initiated, route entry in the
associated virtual server's obj.conf configuration for that
partition will be modified to list the new origin server pool apart
from the existing pool that will handle the new requests. Existing
pool (old or drain pool) will handle only the sticky requests to
its origin servers.
[0087] After this configuration change, for every request to that
partition, OTD will first check if it is a sticky request and if
the origin server which is expected to process the request is part
of the old origin server pool.
[0088] If yes, then the request is routed to the specific origin
server of the old pool. If that old origin server is not active, it
will be routed to another origin server of the old pool.
[0089] If the request is not a sticky one, it is routed to the new
origin server pool.
[0090] The old pool will continue to process the sticky requests
until the configuration is modified to remove the old pool entry.
It will be the admin's responsibility to turn off the draining of
the pool at an appropriate time.
[0091] WLS would invoke the LCM with a second partition migration
REST call specifying the phase: end-migrate which would in turn
invoke the OTD plugin for configuration change in OTD.
Non-Seamless Case
[0092] In accordance with an embodiment, during seamless cases, the
application server will invoke the LCM with a partition update REST
call [0093] //update all partitions associated with the environment
POST/management/lifecycle/latest/environments/{name}/partitions/{id}/upda-
te
[0094] The payload will contain the details of new cluster where
the partition is deployed. The LCM would delegate the call to OTD
Plugin to update the OTD configuration for that partition.
Graceful Shutdown
[0095] In accordance with an embodiment, during this stage, a
graceful shutdown on the resource group is initiated for the
specified (old) target using the resource group shutdown feature.
This would allow handling of inflight work. This will reduce issues
with unprocessed state from singleton services such as JMS
messages, Running jobs, scheduled tasks, and JTA transactions.
Resource containers handle their relative halt/suspend calls in
different ways. For example, the datasource container will do a
full unregister of it's resources from the transaction manager
while the JMS container will not conduct the unregister but will
block further requests/processing of messages. All resource
containers do conduct unregister calls during undeploy. The
unregister behavior of the various resource containers varies,
however, there is always at least a grace period whereby the
container will wait for pending transactions to complete.
Ensuring all JTA Transactions are Complete
[0096] In the common case, when resource group migration occurs,
all transactions involving resources involved will have completed
before migration occurs. In the rare case that all transactions for
a given resource are not complete by the end of the grace period
(this would be the case if, e.g., a database became unavailable for
an extended period of time at the same time resource group
migration was initiated) the resources can still be migrated. This
is due to the fact that resource migration is constrained to a
single domain and the original server can find the resource in it's
new location and conduct recovery.
EJB Stubs
[0097] For clustered EJBs and RMIs, RG migration is accomplished
using the object's replica-aware stub. When a client makes a call
through a replica-aware stub to a service that fails, the stub
detects the failure and retries the call on another replica. Should
the primary server fail, the client's EJB stub automatically
redirects further requests to the secondary application server
instance. At this point, the secondary server creates a new EJB
instance using the replicated state data, and processing continues
on the secondary server. If a client is misdirected (incorrect
primary/secondary server information) the session replication
framework will employ the session state query protocol to determine
the server that is hosting the replica and will send the
information back to the client stub. The client stub will then
redirect its request to the server hosting the replica.
JMS
[0098] If an administrator tries to initiate migration of a
resource group which has JMS config, validation exception would be
thrown. The administrator will be displayed a message to shut down
the resource group before migration. In order to avoid data loss
during RG migration, application messages should be completely
drained and all pending transactions completed before undeploying
or removing a JMS server or SAF Agent. Therefore, it is the job of
the administrator to make sure all the draining is done before
resource group shutdown is initiated. After the resource group
migration is completed, it is the administrator's job to then
restart the resource group.
MBean Configuration Change
[0099] In accordance with an embodiment, when an application stops
or timeout is reached then ResourceGroup's target i.e. virtual
target is updated to remove the old cluster. This is final step of
migration process.
[0100] The present invention may be conveniently implemented using
one or more conventional general purpose or specialized digital
computer, computing device, machine, or microprocessor, including
one or more processors, memory and/or computer readable storage
media programmed according to the teachings of the present
disclosure. Appropriate software coding can readily be prepared by
skilled programmers based on the teachings of the present
disclosure, as will be apparent to those skilled in the software
art.
[0101] In some embodiments, the present invention includes a
computer program product which is a non-transitory storage medium
or computer readable medium (media) having instructions stored
thereon/in which can be used to program a computer to perform any
of the processes of the present invention. The storage medium can
include, but is not limited to, any type of disk including floppy
disks, optical discs, DVD, CD-ROMs, microdrive, and magneto-optical
disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory
devices, magnetic or optical cards, nanosystems (including
molecular memory ICs), or any type of media or device suitable for
storing instructions and/or data.
[0102] The foregoing description of the present invention has been
provided for the purposes of illustration and description. It is
not intended to be exhaustive or to limit the invention to the
precise forms disclosed. Many modifications and variations will be
apparent to the practitioner skilled in the art. The embodiments
were chosen and described in order to best explain the principles
of the invention and its practical application, thereby enabling
others skilled in the art to understand the invention for various
embodiments and with various modifications that are suited to the
particular use contemplated. It is intended that the scope of the
invention be defined by the following claims and their
equivalents.
* * * * *
References